Copyright  2004 by the Genetics Society of America
DOI: 10.1534/genetics.104.026260
A Role for DNA Polymerase ␦ in Gene Conversion and Crossing Over During
Meiosis in Saccharomyces cerevisiae
Laurent Maloisel,*,1 Jaya Bhargava†,2 and G. Shirleen Roeder*,†,‡,3
*Howard Hughes Medical Institute, †Department of Molecular, Cellular and Developmental Biology and
‡
Department of Genetics, Yale University, New Haven, Connecticut 06520-8103
Manuscript received January 6, 2004
Accepted for publication April 15, 2004
ABSTRACT
A screen for mutants of budding yeast defective in meiotic gene conversion identified a novel allele of
the POL3 gene. POL3 encodes the catalytic subunit of DNA polymerase ␦, an essential DNA polymerase
involved in genomic DNA replication. The new allele, pol3-ct, specifies a protein missing the last four
amino acids. pol3-ct shows little or no defect in DNA replication, but displays a reduction in the length
of meiotic gene conversion tracts and a decrease in crossing over. We propose a model in which DNA
synthesis determines the length of strand exchange intermediates and influences their resolution toward
crossing over.
H
OMOLOGOUS recombination plays a critical role
in maintaining genome integrity throughout cell
division. In meiosis, homologous recombination is essential for proper homolog pairing and for the correct
segregation of chromosomes at the first meiotic division.
In vegetative cells, recombination plays an important
role during DNA replication by providing a mechanism
to bypass DNA lesions and other obstacles that block
replication fork progression. Homologous recombination also provides a means to generate new combinations of genetic markers through gene conversion and
crossing over, thereby generating genetic diversity among
different individuals in the same population.
Numerous insights into the mechanism of homologous
recombination have been obtained from meiotic studies
using the convenient model organism, Saccharomyces cerevisiae (Paques and Haber 1999). Meiotic recombination in
budding yeast is initiated by the formation of DNA doublestrand breaks (DSBs) at recombination hotspots. The
strands with 5⬘ ends at the site of the break are processed
to expose single-stranded tails with 3⬘ termini (Figure
1). DSB formation and 5⬘-end resection are followed by
invasion of an intact nonsister chromatid by only one
of the two single-stranded tails (Hunter and Kleckner
2001). Single-end invasion results in hybrid DNA (hDNA)
in which the two strands in a duplex are of different
1
Present address: CEA de Fontenay-aux-Roses, UMR 217 CNRS-CEA/
DSV/DRR/LERA, 92265 Fontenay-aux-Roses, France.
2
Present address: End Stage Renal Disease Network of New England,
30 Hazel Terrace, Woodbridge, CT 06525.
3
Corresponding author: Howard Hughes Medical Institute, Department of Molecular, Cellular and Developmental Biology, Yale University, P.O. Box 208103, New Haven, CT 06520-8103.
E-mail: [email protected]
Genetics 167: 1133–1142 ( July 2004)
parental origin. If the two parental duplexes are genetically different within the region of strand exchange, the
resulting hDNA contains mismatched base pairs and is
referred to as heteroduplex DNA.
Single-end invasion intermediates can be channeled
toward either of two repair pathways. In the first pathway, described by the DSB repair model (Szostak et al.
1983; Sun et al. 1989), DNA synthesis, capture of the
second end, and ligation generate a double Holliday junction intermediate with asymmetric hDNA (i.e., hDNA on
only one of the two duplexes) on each side of the DSB
and on each chromatid (Figure 1, left). If a Holliday
junction undergoes branch migration, then hDNA will
be formed on both duplexes (symmetric hDNA). Eventually, double Holliday junction intermediates are resolved by cutting, at each junction, either both outside
strands or both inside strands. Cutting of the two junctions in opposite directions generates crossovers, while
cutting in the same direction generates noncrossovers.
The second pathway is described by the synthesisdependent strand annealing (SDSA) model (Paques
and Haber 1999) and supported by recent meiotic studies (Allers and Lichten 2001; Merker et al. 2003).
According to the SDSA model (Figure 1, right), the
invading strand is extended by DNA synthesis and then
subsequently displaced. The newly synthesized DNA
strand then anneals to the single-stranded tail on the
other side of the break. DSB repair is completed by
DNA synthesis and ligation. If extension of the invading
strand creates a single-stranded tail that is longer than
the exposed complement on the opposite side of the
break, a flap structure with an exposed 3⬘ end will result
after the two strands anneal and this flap will need to
be removed to allow ligation. In this model, asymmetric
hDNA occurs only on one side of the break and only
1134
L. Maloisel, J. Bhargava and G. S. Roeder
Figure 1.—Models of meiotic recombination. Diagrammed are DSB repair
by the DSB repair pathway (left) and by
SDSA (right). Shown are two recombining DNA duplexes, one indicated in red
and the other in blue. See text for details. In the DSB repair model, arrowheads depict cleavage of double Holliday junctions; only two of four possible
resolution products are shown.
on the chromatid that suffered the DSB. The SDSA
repair pathway does not lead a priori to the formation
of double Holliday junctions and therefore produces
only noncrossover products. Although SDSA is a common mode of DSB repair in vegetative cells (Paques
and Haber 1999), the extent to which it contributes to
meiotic recombination remains unclear.
Although early intermediates in meiotic recombination have been characterized by extensive genetic, biochemical, and physical analyses, the later steps remain
obscure. For example, it is not understood what determines the channeling of single-end invasion intermediates into either of the two repair pathways. In the DSB
repair pathway, it is not known what controls the elongation of strand exchange intermediates, second-end capture, formation of Holliday junctions, branch migration,
or junction resolution. This study provides insight into
some of these issues.
We describe a novel allele of the POL3 gene of S.
cerevisiae coding for the catalytic subunit of DNA polymerase ␦. This mutation has little or no effect on DNA
replication or DNA damage repair in vegetative cells.
However, during meiotic recombination, the mutant
produces shorter strand exchange intermediates and
fewer crossover products. These results lead us to propose a role for Pol␦ in meiotic recombination beyond
an involvement in simple gap repair.
MATERIALS AND METHODS
Isolation and mapping of the pol3-ct mutant: The pol3-ct
mutant was isolated using the homothallic (HO) strain S2702
Polymerase ␦ in Meiotic Gene Conversion
(MATa/MAT␣ and homozygous for ade2-1 HIS4-ura3-Stu-his4260 leu2-3,112 lys2-1 thr1-4 trp1-289 ura3-1). After induction of
sporulation and spore enrichment (Rockmill et al. 1991)
spores were mutagenized with ultraviolet (UV) light to 50%
survival and plated on YPD medium. Spore colonies were
replica plated to sporulation medium and then to medium
lacking uracil to assay the production of uracil prototrophs.
The pol3-ct mutant was identified on the basis of a strong
decrease (ⵑ10-fold) in meiotic Ura⫹ prototroph formation.
Mapping of the pol3-ct mutation was carried out using multiply marked strains (X4119-15D, STX145-15D, STX82-3A,
STX153-10C, STX75-3C, STX147-4C, X4120-19D, STX-6C,
and STX155-9B) obtained from the Yeast Genetic Stock Center. Strains from our own collection carrying mutations in
various meiotic genes (e.g., MSH5 and REC102) were also used.
Strains and plasmids: Strains used to characterize the effect
of pol3-ct in meiosis are derived from the haploid strains AS4
(MAT␣ trp1-1 arg4-17 tyr7-1 ade6 ura3 MAL2) and AS13 (MATa
leu2-Bst ura3 ade6 ) (Nag et al. 1989). AS4 is a HIS4 strain,
while the PD strains are his4 strains derived from AS13 by
transformation. PD75 is his4-ACG (a T to C transition in the
initiating codon), PD5 is his4-519 (a 1-bp insert at ⫹473),
PD22 is his4-712 (a 1-bp insert at ⫹1396), PD24 is his4-713 (a
1-bp insert at ⫹2270), and DNY25 is his4-lopc (insertion of a
palindromic sequence at the Sal I site in HIS4). pol3-ct derivatives of these strains were constructed by the two-step transplacement procedure (Rothstein 1991). In the first step,
strains were transformed with HindIII-digested LM␦1 (see below). Then Ura⫺ derivatives of Ura⫹ transformants were selected on 5-fluoroorotic acid, creating strains LMP6 (AS4),
LMP5 (PD75), LMP2 (PD5), LMP3 (PD22), LMP4 (PD24),
and LMP8 (DNY25). Replacement of the POL3 allele by pol3ct was verified by DNA sequencing.
Plasmid GR160 (obtained from Roland Chanet) containing
the wild-type POL3 gene in YCp50 was used to complement
the pol3-ct defect. The POL3 gene is toxic in many constructs in
Escherichia coli (R. Chanet, personal communication), which
probably accounts for our inability to clone POL3 on the basis
of complementation of the mutant defect.
Plasmid LM␦1 was designed to allow replacement of the
wild-type copy of POL3 at the endogenous locus by the pol3ct allele. The 3⬘ end of the POL3 gene was amplified from a
pol3-ct strain using primers p2862 (TAGTAGGAATTCTTGCT
TCTGTCCGTCGTGA) and pR4173 (TAGAAGTCGACTAGC
GCCCGAAGTCCTCACA). p2862 anneals starting at position
⫹2503 within POL3, while pR4173 anneals downstream of
POL3. The amplified fragment is 1311 bp in length. Underlined nucleotides within primers indicate restriction sites for
EcoRI (p2862) and Sal I (pR4173) that have been added at
the 5⬘ end of the primers to clone the amplified fragment
into plasmid pRS306 (New England Biolabs, Beverly, MA)
treated with EcoRI and Sal I. The HindIII site used for targeting
is located 260 bp from the Sal I site.
All phenotypic characterizations of pol3-ct were performed
in the AS strain background except for mitotic prototroph
formation, which was assayed in both AS strains and strain
BR2495 (Rockmill and Roeder 1990) and its pol3-ct derivative
LMR1/2. BR2495 has the following genotype: MATa/MAT␣
his4-280/his4-260 leu2-27/leu2-3,112 arg4-8/ARG4 thr1-1/thr1-4
cyh10/CYH10 ade2-1/ade2-1 ura3-1/ura3-1 trp1-1/trp1-289.
Forward mutation, mitotic recombination, and irradiation
assays: Forward mutation to canavanine resistance and mitotic
intragenic recombination were determined by fluctuation
tests using the method of the median. Rates reported are the
average of three (30⬚) or two (18⬚) independent experiments,
each performed with nine independent 5-ml cultures set up
from 2-day-old colonies and incubated at 30⬚ or 18⬚.
Cells in stationary phase (UV) or exponentially growing
(␥-rays) were washed in 0.9% NaCl and plated at appropriate
1135
dilutions on YPD and synthetic medium. UV irradiation was
performed using a 264-nm source delivering 1 J/m2/sec. ␥
irradiation was performed using a 137Cs source. Cells were
treated with 0, 100, 200, and 400 Gy. Survival was determined
as the number of cells forming colonies on YPD medium after
a given dose of irradiation divided by the number of colonies
derived from cells not irradiated. Each experiment was carried
out at least three times with qualitatively similar results.
Genetic analysis: For meiotic analysis of AS strains, cells
were incubated in 5 ml of liquid YPD medium overnight,
washed once with water, and then resuspended in 10 ml liquid
sporulation medium and incubated at 18⬚ for 5 days. Tetrads
were treated with zymolyase in 1 m sorbitol for 10 min and
dissected on YPAD plates. After 3 days, they were replica plated
to appropriate omission media to follow the segregation of nutritive markers. Data were used from four-spore-viable tetrads only,
which composed ⵑ60% of tetrads for both wild-type and pol3-ct
strains. To compare nonparental ditype (NPD) ratios between
wild-type and pol3-ct strains, a contingency chi-square test was
performed for each interval, using the observed and expected
NPD values for wild type and mutant.
RESULTS
Isolation and genetic mapping of the pol3-ct mutant:
A novel screen for mutants defective in meiotic gene
conversion was devised. The starting strain is a HO diploid homozygous for the ura3-1 allele at the URA3 locus
on chromosome V and for a ura3-Stu mutation inserted
at the HIS4 locus on chromosome III. This strain was
sporulated, and the spores were mutagenized with UV
light to 50% survival. Due to mating-type switching followed by mating between cells of opposite mating type,
the cells in spore colonies are diploid and homozygous
at all loci except MAT. Since any newly induced mutations are homozygous, recessive mutations can be
readily detected. The diploids still carry ura3-1 and ura3Stu alleles and thus can be assayed for meiotic gene
conversion by screening for the production of Ura⫹
prototrophs. A conversion-defective mutant that showed
a 10-fold reduction in Ura⫹ prototrophs was chosen for
further study.
Attempts to clone the gene from a yeast genomic
library by complementation of the conversion defect
proved unsuccessful. We therefore mapped the gene by
tetrad analysis. The mutant was crossed to ura3-1 strains
carrying 48 different markers dispersed throughout the
yeast genome (Figure 2). Haploid segregants were
scored for a defect in meiotic gene conversion by mating
to tester strains carrying ura3-Stu and the conversiondefective allele; the resulting diploids were sporulated
and screened for the production of uracil prototrophs.
This mapping localized the mutation to the left arm of
chromosome IV, 22 cM centromere proximal to MSH5
and 30 cM distal to MBP1.
In the region defined by mapping, the POL3 gene,
which encodes the catalytic subunit of DNA polymerase
␦ (Pol␦), was the only obvious candidate for the mutated
gene. Consistent with this possibility, a plasmid carrying
a wild-type copy of POL3 complements the mutant defect. When the POL3 gene was amplified from the mu-
1136
L. Maloisel, J. Bhargava and G. S. Roeder
Figure 2.—Mapping of conversion-defective
mutant. A genetic map showing the 16 chromosomes of yeast is diagrammed. Centromeres are
represented by white ovals. The pol3-ct mutation
was mapped by crossing to strains carrying 48
different markers dispersed throughout the genome. Beyond a genetic distance of 50 cM, genetic independence is expected. Therefore, tetrad analyses showing no linkage between the
mutation of interest and a given marker define a
region of 50 cM on both sides of the marker in
which the mutation cannot lie. Such regions are
indicated by rectangles, shown in black up to 35
cM away from the marker and in gray from 35
cM to 50 cM (e.g., the ADE3 marker on chromosome VII). Centromeric markers (e.g., ADE1 and
TRP1) indicated that the conversion-defective
mutation is not linked to its centromere; therefore, all centromeres are covered by rectangles.
Mapping showed that the pol3-ct mutation is approximately at the center of a region defined by
the MSH5 and MBP1 genes.
tant strain by PCR and sequenced, a point mutation
was found in the POL3 open reading frame at position
⫹3268, 11 bp before the last nucleotide (Figure 3A).
This allele changes a leucine codon (TTA) to a stop
codon (TAA). As a consequence, the protein is missing
the last four amino acids (LSKW). The mutant was therefore named pol3-ct, for POL3 C-terminal truncation. The
mutation does not lie in any of the conserved domains
of the protein; in particular, pol3-ct does not affect the
putative zinc finger domains also located near the carboxy terminus of the protein.
pol3-ct does not show growth or DNA repair defects:
Since POL3 is an essential gene responsible for the polymerase and proofreading activities of Pol␦ (Hindges
and Hubscher 1997), it was important to determine
the effects of pol3-ct in vegetative cells. Growth of pol3ct mutants is normal at 18⬚ and 30⬚; pol3-ct grows slightly
slower than wild type at 37⬚ (doubling time of 150 min
for POL3 vs. 165 min for pol3-ct). These results indicate
that DNA synthesis is not impaired in pol3-ct mutants as
it is in thermosensitive mutants of POL3 that show cell
cycle arrest when switched to nonpermissive temperatures (Conrad and Newlon 1983).
Replication defects caused by a deficient Pol␦ can
result in the accumulation of mutations (Datta et al.
2000) and/or DNA lesions that are potential recombination substrates (Aguilera and Klein 1988). In addition,
Pol␦ has been proposed to play a role in DNA damage
repair (Giot et al. 1997). We therefore examined the
effect of pol3-ct on mutation rate, mitotic recombination, and sensitivity to DNA-damaging agents. These
experiments were carried out at 30⬚, the optimum temperature for yeast cell growth. In addition, since the
meiotic experiments described below were carried out
at 18⬚, a subset of vegetative phenotypes was also examined at this temperature. We found no change in mutation rate at the CAN1 locus at 30⬚ and a modest (but
statistically significant) twofold increase at 18⬚ (Figure
3B). There was no change in mitotic intragenic recombination at either temperature (Figure 3, C and D). The pol3ct mutant does not display enhanced sensitivity to UV irradiation (Figure 3, E and F) or to ␥ irradiation (Figure 3G).
pol3-ct changes the conversion gradient at HIS4: The
pol3-ct allele was originally found to decrease meiotic
intragenic recombination between heteroalleles of the
URA3 gene, suggesting that pol3-ct affects gene conver-
Polymerase ␦ in Meiotic Gene Conversion
1137
Figure 3.—Characterization of pol3-ct mutant. (A)
Pol3 protein and location of
the pol3-ct mutation. The
coding sequence includes
six regions involved in polymerase activity (domains
I–VI, shaded), three regions
corresponding to the exonuclease proofreading active site (Exo I, II, and III,
shaded; Exo II is contained
within domain IV), and a
putative zinc finger DNAbinding domain (Zn; Hindges and Hubscher 1997).
The position of the pol3-ct
mutation is indicated by an
arrow. (B) Mutation rates at
the CAN1 locus determined
at 18⬚ and 30⬚. (C) Mitotic intragenic recombination rates
between HIS4 heteroalleles
determined at 18⬚ and 30⬚.
(D) Mitotic intragenic recombination rates at 30⬚ determined in strain BR2495
carrying heteroalleles at
four different loci. Error
bars in B–D represent standard deviations. (E and F)
Sensitivity to UV light at 18⬚
and at 30⬚. (G) Sensitivity to
␥ irradiation at 30⬚. In E–G,
solid circles represent diploid
strains, and open circles indicate haploid strains; solid
lines represent wild-type
strains, and dashed lines
represent pol3-ct strains. A
rad52 haploid, known to be
highly sensitive to ␥-rays,
was used as a control
(square symbols).
sion. To determine which step(s) of meiotic gene conversion might be altered, conversion was examined at the
HIS4 recombination hotspot (Nag et al. 1989). These experiments were carried out at 18⬚, the temperature at
which the frequency of non-Mendelian segregation is maximal in the strain background used for this analysis.
Non-Mendelian segregation is observed at high frequency at HIS4 (Table 1) due to the presence of a
meiotic DSB site just upstream of the gene (Fan et al.
1995). hDNAs spanning regions of heterologies within
HIS4 contain mismatches. Correction of these mismatches can lead to gene conversion (3:1 or 1:3 segregations), whereas unrepaired mismatches lead to postmeiotic segregations (sectored colonies). An important
feature of the HIS4 meiotic hotspot is that it shows a
conversion gradient with its high end at the 5⬘ end of
the gene next to the initiation site (Detloff et al. 1992).
Remarkably, the conversion gradient within HIS4 is
strongly accentuated in the pol3-ct background (Table
1; Figure 4A). No effect on non-Mendelian segregation
frequencies was seen for the most 5⬘ marker, indicating
that initiation of recombination at HIS4 is not altered
in a pol3-ct background. In contrast, there is a decrease
in gene conversion for all other markers in HIS4. The
effect of pol3-ct increases progressively within HIS4 to a
maximum of sixfold at the 3⬘ end of the gene. We therefore conclude that formation of hDNAs starting at the
initiation site upstream of HIS4 is not impaired in the
pol3-ct mutant. However, these hDNAs extend shorter
distances into HIS4 in pol3-ct than in wild type.
Previous studies have shown that polarity gradients
are largely abolished when mismatch repair is inhibited
(Detloff et al. 1992; Alani et al. 1994). Two observations indicate that the effect of pol3-ct on the HIS4 conversion gradient is not due to an alteration in the efficiency of mismatch repair. First, for none of the his4
1138
L. Maloisel, J. Bhargava and G. S. Roeder
TABLE 1
Effect of the pol3-ct mutation on non-Mendelian segregation frequencies of heterozygous markers
Diploid
strain
AS4/PD75
LMP6/LMP5
AS4/PD5
LMP6/LMP2
AS4/PD22
LMP6/LMP3
AS4/PD24
LMP6/LMP4
AS4/DNY25
LMP6/LMP8
AS/PD
LMP/LMP
AS/PD
LMP/LMP
AS/PD
LMP/LMP
AS/PD
LMP/LMP
Marker
his4-ACG
his4-ACG
his4-519
his4-519
his4-712
his4-712
his4-713
his4-713
his4-lopc
his4-lopc
arg4-17
arg4-17
tyr7-1
tyr7-1
leu2-Bst
leu2-Bst
MAT
MAT
Position
in HIS4
⫹2
⫹2
⫹473
⫹473
⫹1396
⫹1396
⫹2270
⫹2270
⫹473
⫹473
Segregation pattern
POL3
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
wt
pol3-ct
4:4
6:2
251 88
243 80
296 53
376 33
378 33
458
9
388 19
344
1
247 41
360 28
2062 101
2076 35
975 17
1210 10
2184 21
2079 24
2197 27
2136
6
2:6 5:3 3:5
98
95
65
38
70
14
44
7
11
4
85
33
25
19
43
41
24
2
5
0
0
0
0
0
1
0
62
16
0
0
6
1
0
0
0
0
5
0
0
0
0
0
0
0
45
20
0
0
0
0
0
0
0
0
ab4:4
0
0
0
0
0
0
0
0
12
3
0
0
0
0
0
0
0
0
Total
Fold
8:0 0:8 Other tetrads % NMS decrease
6
8
3
1
1
0
1
0
1
0
1
0
1
0
0
0
0
0
10
4
1
1
1
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5
1
0
0
0
0
0
0
0
0
463
430
425
449
483
481
452
352
425
432
2248
2144
1024
1240
2248
2144
2248
2144
45.7
43.6
30.3
16.2
21.7
4.8
14.1
2.3
41.9
16.6
8.4
3.1
4.8
2.4
2.8
3.0
2.3
0.4
1.0
1.9*
4.5*
6.1*
2.5*
2.7*
2.0*
1.0
5.7*
Diploid strains are designated according to the names of their haploid parents (see materials and methods). Data for the
ARG4, TYR7, LEU2, and MAT markers represent pooled data from the five wild-type strains and the five pol3-ct strains used to
examine the segregation of different HIS4 alleles. wt, POL3/POL3 diploid; pol3-ct, pol3-ct/pol3-ct diploid. The nomenclature used
for meiotic segregation frequencies refers to the eight DNA strands (four chromatids) present in meiotic prophase. The different
types of meiotic segregation are 4:4 (normal Mendelian segregation), 6:2 and 2:6 (gene conversion), 5:3 and 3:5 (tetrads with
one sectored colony indicating an unrepaired mismatch in hDNA), ab4:4 (aberrant 4:4; one wild-type, one mutant, and two
sectored colonies), and 8:0 and 0:8 (tetrads with four spores of a single genotype). “Other” includes aberrant 6:2 segregations
(two wild-type and two sectored colonies) and aberrant 2:6 segregations (two mutant and two sectored colonies) as well as 7:1
segregations (one sectored and three wild-type colonies) and 1:7 (one sectored and three mutant colonies). The percentages
of non-Mendelian segregation in wild-type and pol3-ct strains were compared by chi-square tests. *Fold decreases from wild type
that are statistically significant (P ⬍ 0.001). NMS, non-Mendelian segregation.
alleles examined does the pol3-ct mutation increase the
frequency of 5:3 and 3:5 segregations (indicative of a
failure to repair) relative to 6:2 and 2:6 segregations
(indicative of repair). Second, a HIS4 mutation (his4-lopc)
that usually escapes mismatch repair (even in wild type)
shows a decreased frequency of non-Mendelian segregation in the pol3-ct mutant (Table 1).
The pol3-ct effect on gene conversion is not specific
to the HIS4 locus: The decrease in meiotic gene conversion in pol3-ct is not specific to HIS4. A twofold decrease
in conversion was also found at the ARG4 and TYR7 loci
(Table 1). At the MAT locus, parental alleles in diploids
differ by a heterology of 700 bp. An mlh1 mutant, which
is deficient in mismatch repair, shows high frequencies
Figure 4.—Gene conversion, crossing over, and crossover interference in the pol3ct mutant. (A) The solid lines
indicate the frequency of
non-Mendelian segregation
at different positions in HIS4
for wild-type and pol3-ct
strains. The dashed line indicates the fold decrease in nonMendelian segregation frequency in pol3-ct compared to
wild type. The graph is derived from the data in Table 1. (B) Map distances in centimorgans for four different intervals in wild type and pol3-ct. (C) Crossover
interference in the HIS4-LEU2 and LEU2-MAT intervals. The dashed line indicates the NPD value expected in the absence of interference.
Histograms in B and C are derived from the data in Table 2.
Polymerase ␦ in Meiotic Gene Conversion
1139
TABLE 2
Effect of the pol3-ct mutation on map distances and NPD ratios
Interval
Genotype
HIS4-LEU2
HIS4-LEU2
LEU2-MAT
LEU2-MAT
POL3
pol3-ct
POL3
pol3-ct
Interval
Genotype
ARG4-CENVIII
ARG4-CENVIII
TYR7-CENXVI
TYR7-CENXVI
POL3
pol3-ct
POL3
pol3-ct
PD
NPD
TT
cM
866
1197
1038
1210
29
13
66
22
615
512
1006
835
26.1
17.1
33.2
23.4
PD ⫹ NPD
TT
cM
% wt
1552
1721
409
623
470
196
565
587
11.6
5.1
29
24.2
% wt
Prob.
65
⬍0.001
70
⬍0.001
NPD
exp.
NPD
ratio
Prob.
45
24
95
60
0.64
0.54
0.69
0.37
⬍0.02
⬍0.05
⬍0.01
⬍0.001
Prob.
44
⬍0.001
83
⬍0.001
For each interval, the size of the map distance as a percentage of the map distance in wild type (% wt) is indicated. Also
presented is the probability (Prob.), based on a chi-square test, that the difference between wild-type and mutant map distances
is due to chance. For the HIS4-LEU2 and LEU2-MAT intervals, map distances were calculated using the following formula: map
distance (in centimorgans) ⫽ [(TT ⫹ 6NPD)/2(PD ⫹ NPD ⫹ TT)] ⫻ 100. The number of NPD tetrads expected (NPD exp.)
in the absence of interference was calculated on the basis of the observed number of TT tetrads (Papazian 1952). The NPD
ratio is the number of NPD tetrads observed divided by the number expected. The probability (Prob.), based on a chi-square
test, that the difference between the observed and expected numbers of NPD tetrads is due to chance is indicated. Map distance
between ARG4 (or TYR7) and its centromere was determined using a trp1 mutation, which is tightly linked to its centromere
on chromosome IV. In this case, map distances were calculated using the following formula: map distance (in centimorgans) ⫽
[1⁄2TT/(PD ⫹ NPD ⫹ TT)] ⫻ 100. PD, parental ditype; NPD, nonparental ditype; TT, tetratype.
of postmeiotic segregation at MAT, indicating formation of hDNA with large loops at this locus (Wang et
al. 1999). Conversion of this heterology is decreased
sixfold in pol3-ct (Table 1), strengthening the conclusion
that the decrease in gene conversion does not depend
on the nature of the segregating mutations. No decrease
in conversion was observed for the leu2-Bst marker. On
the basis of the results obtained at HIS4, it is likely that
the leu2-Bst marker is close to a recombination initiation
site.
Crossing over is decreased in pol3-ct strains: pol3-ct
strains show wild-type levels of sporulation (ⵑ30%) and
spore viability (ⵑ90%). These results, together with the
studies of gene conversion described above, indicate
that DSB formation and repair occur as efficiently in
pol3-ct strains as they do in wild type. To gain insight
into the resolution of recombination intermediates in
the mutant, crossover formation was analyzed by measuring map distances in four genetic intervals on three
different chromosomes. In all intervals, map distances
are decreased, on average to 66% of the wild-type level
(Table 2; Figure 4B). These results suggest a role for
DNA synthesis in promoting crossover formation.
Crossovers show interference in pol3-ct strains: Meiotic crossovers are nonrandomly distributed such that
two crossovers rarely occur close together, a phenomenon known as crossover interference. The effect of the
pol3-ct mutation on interference was investigated by measuring NPD ratios, which can be roughly defined as the
frequency of double crossovers observed in a marked
interval divided by the number expected in the absence
of interference (Papazian 1952; Snow 1979). An NPD
ratio of ⬍1.0 indicates positive interference. Crossover
interference is not decreased in the pol3-ct mutant (Table 2; Figure 4C). In fact, in both intervals tested, the
NPD ratio in the mutant is lower than that in wild type
(indicating stronger interference); this difference is statistically significant only in the LEU2-MAT interval (P ⬍
0.05).
DISCUSSION
pol3-ct is a separation-of-function allele: Previous studies of the role of Pol␦ in recombination have been complicated by the use of temperature-sensitive alleles,
which have modest defects in DNA replication even
at permissive temperatures. Nevertheless, it has been
possible to show that Pol␦ is required in vegetative cells
for gene conversion events induced by irradiation and
for repair of a DSB at the MAT locus (Fabre et al. 1991;
Holmes and Haber 1999). This study is the first to
demonstrate a role for Pol␦ in meiotic recombination.
Our data demonstrate that the pol3-ct mutant has little
or no defect in DNA replication or DNA damage repair
in vegetative cells. In addition, we have found that pol3ct does not affect spore formation or spore viability,
indicating that premeiotic DNA replication is efficient.
Thus, pol3-ct acts as a separation-of-function allele that
specifically impairs a function of Pol␦ involved in meiotic recombination. This characteristic of the pol3-ct allele makes it a unique tool to investigate the role of Pol␦
1140
L. Maloisel, J. Bhargava and G. S. Roeder
Figure 5.—Models for the effect of the pol3-ct mutation on
meiotic recombination. (A) Proposed effect of pol3-ct in the
SDSA pathway. (B) Proposed effect of pol3-ct on the DSB repair
pathway. See text for explanation.
in meiotic recombination, unimpeded by the complications associated with temperature-sensitive mutations.
Possible roles for Pol␦ in determining gene conversion tract length: Our results suggest that pol3-ct strains
initiate the wild-type number of meiotic recombination
events; however, the average length of hDNA formed
is shorter than that in wild type. It is not clear why the
pol3-ct mutation was originally identified on the basis of
decreased prototroph formation between URA3 heteroalleles. Perhaps both URA3 alleles are far from a DSB
site; in this case, shortening tract length might decrease
the fraction of events that reach either mutation.
What does the decrease in gene conversion tract
length in pol3-ct strains tell us about the role of wild-type
Pol␦ in meiotic recombination? Recombination events
channeled through the DSB repair pathway necessitate
the synthesis of new DNA sufficient to replace the sequences removed by processing of DSBs to expose single-stranded tails (Figure 1, left). Therefore, in this pathway, the length of asymmetric hDNA should be at least
equivalent to the lengths of the single-stranded tails
present at DSB sites. In contrast, in the SDSA pathway,
displacement of the invading strand might occur before
DNA synthesis extends the full length of the singlestranded tail on the other side of the DSB site (Figure
5A). In this case, repair synthesis would have to be completed after annealing of the displaced strand to the
complementary single strand. The length of hDNA
would then be confined to the length of DNA synthesis
that occurred on the invaded chromatid. In the pol3ct mutant, DNA repair synthesis might frequently stop
earlier than in wild type, perhaps due to weakened processivity of Pol␦ during recombination.
How might the pol3-ct mutation affect tract length for
gene conversion events that proceed through the DSB
repair pathway? We propose a model in which Pol␦ does
not simply serve to fill in single-stranded gaps, but rather
plays a determining role in hDNA extension. According
to the DSB repair model, DNA synthesis primed from
the 3⬘ end of the invading strand enlarges the D-loop;
the displaced DNA subsequently captures the second
single-stranded tail, which in turn primes repair synthesis. We propose that Pol␦ subsequently increases the
length of asymmetric hDNA by promoting DNA synthesis that is coupled to extended removal of the resected
strand either by strand displacement or by further 5⬘ to
3⬘ resection (Figure 5B). In this scenario, DNA synthesis
would be de facto coupled with junction migration; consequently, asymmetric hDNA could be extended prior
to the formation of ligated double Holliday junctions.
This hypothesis would explain nicely why asymmetric
hDNAs at HIS4 (and at other recombination hotspots
such as ARG4 and CYS3) frequently span the entire gene
while the initial single-stranded tails appear to be no
longer than 800 nucleotides (Sun et al. 1991; Bishop et
al. 1992; Detloff and Petes 1992; Vedel and Nicolas
1999).
A specific interaction between Pol␦ and a component
of the recombination apparatus might be required for
synthesis-coupled junction migration. In the pol3-ct mutant, this interaction might not occur efficiently such
that Pol␦ is removed upon reaching the first 5⬘ end,
thus preventing further hDNA extension.
Possible roles for Pol␦ in meiotic crossing over: In
addition to altering the length of gene conversion tracts,
pol3-ct also decreases the frequency of crossing over.
How can this observation be reconciled with the models
just presented (Figure 5)?
If single-end invasion intermediates sustain shorter
tracts of DNA synthesis in pol3-ct (Figure 5B), this would
facilitate strand displacement and favor the repair of
DSBs through the SDSA pathway. Since SDSA is not
associated with crossing over, a bias toward SDSA (vs.
the DSB repair pathway) could account for the decrease
in crossing in pol3-ct strains.
An alternative (not mutually exclusive) possibility is
that pol3-ct decreases the probability of crossing over for
events that proceed through the DSB repair pathway.
Pol␦ might be involved in isomerization and/or the
decision as to which strands are to be cut during Holliday junction resolution. For example, the resolvase that
introduces nicks into strands of like polarity at Holliday
junctions could be directed by Pol␦ to operate on
strands with newly synthesized DNA. Such a bias would
favor cutting the two junctions in opposite directions
and thereby encourage crossover formation. Premature
removal of Pol␦ would mean loss of the bias and a
decrease in crossing over.
pol3-ct does not impair crossover interference: A number of mutants (e.g., zip1, msh4) show a two- to threefold
decrease in crossing over that is not associated with decreased initiation of recombination (Sym and Roeder
Polymerase ␦ in Meiotic Gene Conversion
1994; Novak et al. 2001). These mutations also eliminate
crossover interference. The pol3-ct mutation reduces
crossing over without decreasing interference, raising
the possibility that pol3-ct affects a different subset of
crossovers from those eliminated by zip1 and msh4. The
fact that interference appears slightly stronger in the
pol3-ct mutant raises the possibility that pol3-ct specifically
(or preferentially) reduces a subset of crossovers that
do not normally exhibit interference. The existence of
two crossover pathways, one subject to interference and
the other free of interference, is consistent with recent
observations (de los Santos et al. 2003).
Meiotic crossovers are nonrandomly distributed
among chromosomes, such that every chromosome pair
almost always undergoes at least one crossover (an obligate crossover) to ensure its correct segregation at meiosis I. Crossover interference and obligate crossing over
are often assumed to be mechanistically related. Consistent with this hypothesis, mutations that decrease crossover interference also reduce spore viability due to the
missegregation of nonrecombinant chromosomes (Sym
and Roeder 1994; Chua and Roeder 1997; Novak et
al. 2001). The wild-type level of spore viability observed
in pol3-ct strains suggests that crossovers are properly
distributed among chromosomes.
The pol3-ct mutation links replication and recombination: Further studies using the pol3-ct mutant will lead
to a better understanding of the mechanism by which
Pol␦ influences gene conversion tract length and crossing over. This study casts new light on our growing awareness of the links between DNA replication and homologous recombination. It is now generally acknowledged
that homologous recombination serves as a backup system supporting DNA replication when the template
DNA is damaged or the replication machinery malfunctions (Kuzminov 2001). It is fascinating to unveil, on
the other hand, that the outcome of homologous recombination depends largely on a major DNA replication
protein.
We thank Roland Chanet and Tom Petes for kindly providing plasmids and yeast strains. L.M. especially thanks Eun-Jin Erica Hong for
valuable comments on the manuscript and Michael Odell, Roberta
Shew, and Nadia Shapiro for excellent technical assistance. This work
was supported by the Howard Hughes Medical Institute, le Commissariat à l’Energie Atomique, and Electricité de France. L.M. was supported by postdoctoral fellowships from the Association pour la Recherche sur le Cancer and from the European Molecular Biology
Organization.
LITERATURE CITED
Aguilera, A., and H. L. Klein, 1988 Genetic control of intrachromosomal recombination in Saccharomyces cerevisiae. Isolation and
genetic characterization of hyper-recombination mutations. Genetics 119: 779–790.
Alani, E., R. A. G. Reenan and R. D. Kolodner, 1994 Interaction
between mismatch repair and genetic recombination in Saccharomyces cerevisiae. Genetics 137: 19–39.
1141
Allers, T., and M. Lichten, 2001 Differential timing and control
of noncrossover and crossover recombination during meiosis.
Cell 106: 47–57.
Bishop, D. K., D. Park, L. Xu and N. Kleckner, 1992 DMC1: a
meiosis-specific yeast homolog of E. coli recA required for recombination, synaptonemal complex formation, and cell cycle progression. Cell 69: 439–456.
Chua, P. R., and G. S. Roeder, 1997 Tam1, a telomere-associated
meiotic protein, functions in chromosome synapsis and crossover
interference. Genes Dev. 11: 1786–1800.
Conrad, M. N., and C. S. Newlon, 1983 Saccharomyces cerevisiae cdc2
mutant fails to replicate approximately one-third of their nuclear
genome. Mol. Cell. Biol. 3: 1000–1012.
Datta, A., J. L. Schmeits, N. S. Amin, P. J. Lau, K. Muyung et al.,
2000 Checkpoint-dependent activation of mutagenic repair in
Saccharomyces cerevisiae pol3-01 mutants. Mol. Cell 6: 593–603.
de los Santos, T., N. Hunter, C. Lee, B. Larkin, J. Loidl et al., 2003
The Mus81/Mms4 endonuclease acts independently of doubleHolliday junction resolution to promote a distinct subset of crossovers during meiosis in budding yeast. Genetics 164: 81–94.
Detloff, P., and T. D. Petes, 1992 Measurements of excision repair
tracts formed during meiotic recombination in Saccharomyces cerevisiae. Mol. Cell. Biol. 12: 1805–1814.
Detloff, P., M. A. White and T. D. Petes, 1992 Analysis of a gene
conversion gradient at the HIS4 locus in Saccharomyces cerevisiae.
Genetics 132: 113–123.
Fabre, F., A. Boulet and G. Faye, 1991 Possible involvement of
the yeast POLIII DNA polymerase in induced gene conversion.
Mol. Gen. Genet. 229: 353–356.
Fan, Q., F. Xu and T. D. Petes, 1995 Meiosis-specific double-strand
DNA breaks at the HIS4 recombination hotspot in the yeast Saccharomyces cerevisiae : control in cis and trans. Mol. Cell. Biol. 15:
1679–1688.
Giot, L., R. Chanet, M. Simon, C. Facca and G. Faye, 1997 Involvement of the yeast DNA polymerase ␦ in DNA repair in vivo.
Genetics 145: 1239–1251.
Hindges, R., and U. Hubscher, 1997 DNA polymerase ␦, an essential enzyme for DNA transactions. Biol. Chem. 378: 345–362.
Holmes, A. M., and J. E. Haber, 1999 Double-strand break repair in
yeast requires both leading and lagging strand DNA polymerases.
Cell 96: 415–424.
Hunter, N., and N. Kleckner, 2001 The single-end invasion: an
asymmetric intermediate at the double-strand break to doubleHolliday junction transition of meiotic recombination. Cell 106:
59–70.
Kuzminov, A., 2001 DNA replication meets genetic exchange: chromosomal damage and its repair by homologous recombination.
Proc. Natl. Acad. Sci. USA 98: 8461–8468.
Merker, J. D., M. Dominska and T. D. Petes, 2003 Patterns of
heteroduplex formation associated with the initiation of meiotic
recombination in the yeast Saccharomyces cerevisiae. Genetics 165:
47–63.
Nag, D. K., M. A. White and T. D. Petes, 1989 Palindromic sequences in heteroduplex repair in yeast. Nature 340: 318–320.
Novak, J. E., P. Ross-Macdonald and G. S. Roeder, 2001 The
budding yeast Msh4 protein functions in chromosome synapsis
and the regulation of crossover distribution. Genetics 158: 1013–
1025.
Papazian, H. P., 1952 The analysis of tetrad data. Genetics 37: 175–
188.
Paques, F., and J. E. Haber, 1999 Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.
Microbiol. Mol. Biol. Rev. 63: 349–404.
Rockmill, B., and G. S. Roeder, 1990 Meiosis in asynaptic yeast.
Genetics 126: 563–574.
Rockmill, B., E. J. Lambie and G. S. Roeder, 1991 Spore enrichment. Methods Enzymol. 194: 146–149.
Rothstein, R., 1991 Targeting, disruption, replacement and allele
rescue: integrative DNA transformation in yeast. Methods Enzymol. 194: 281–301.
Snow, R., 1979 Maximum likelihood estimation of linkage and interference from tetrad data. Genetics 92: 231–245.
Sun, H., D. Treco, N. P. Schultes and J. W. Szostak, 1989 Doublestrand breaks at an initiation site for meiotic gene conversion.
Nature 338: 87–90.
Sun, H., D. Treco and J. W. Szostak, 1991 Extensive 3⬘-overhang-
1142
L. Maloisel, J. Bhargava and G. S. Roeder
ing, single-stranded DNA associated with the meiosis-specific double-strand breaks at the ARG4 recombination initiation site. Cell
64: 1155–1161.
Sym, M., and G. S. Roeder, 1994 Crossover interference is abolished
in the absence of a synaptonemal complex protein. Cell 79: 283–
292.
Szostak, J. W., T. L. Orr-Weaver, R. J. Rothstein and F. W. Stahl,
1983 The double-strand-break repair model for recombination.
Cell 33: 25–35.
Vedel, M., and A. Nicolas, 1999 CYS3, a hotspot of meiotic recombi-
nation in Saccharomyces cerevisiae : effects of heterozygosity and
mismatch repair functions on gene conversion and recombination intermediates. Genetics 151: 1245–1259.
Wang, T.-F., N. Kleckner and N. Hunter, 1999 Functional specificity of MutL homologs in yeast: evidence for three Mlh1-based
heterocomplexes with distinct roles during meiosis in recombination and mismatch correction. Proc. Natl. Acad. Sci. USA 96:
13914–13919.
Communicating editor: L. S. Symington